36 BULLETIN 1440, IT. S. DEPARTMENT OF AGRICULTURE 
For this present study, therefore, the coefficients of determina- 
tion, showing the relative significance of each variable over the period 
studied and for the particular combination of variables included, may 
be used as indicating only in a very general way the relative signifi- 
cance of each of the independent variables as factors related to hog 
prices. They would be modified by the inclusion of other inde- 
pendent factors, hence can not be interpreted as measures of absolute 
relationships. 
For the next step in getting at the demand curve, the hog prices 
were deflated; that is, divided by the Bureau of Labor all-commodity 
index of wholesale prices (17). This adjusts the prices to a basis of 
constant " purchasing power," and leaves any further relation 
between hog prices and the price level to show up in the relation to 
the index of business cycles and other related factors. 
At the same time, factors X 2 , slaughter for the previous six months, 
and X 4 , stocks in storage, were dropped from the factors considered. 
Storage was left out for the reason already given — that with it in, 
much of the effect of supply upon price showed up by way of the 
storage regression — and the slaughter for the previous six months 
was left out for the reason that it was more closely related to storage 
stocks than to the moving average of slaughter 17 and would have 
reflected much of the effects of storage had the latter alone been left 
out. 
Dropping out the three factors, X 2 , slaughter for six months 
previous, X 4 , storage stocks, and X !0 , the price index, left seven 
independent variables to be correlated with the new dependent 
variable — hog price divided by the index of wholesale prices — which 
will hereafter be designated X i2 . The multiple correlation of the 
seven independent variables with the new dependent gave R = 0.862, 
or an effective 11 = 0.850, after correcting for the number of variables 
considered. While decidedly lower than the previous corrected 
R of 0.928, this is still fairly close, and serves to give an approximate 
measure of the demand curve. 
The regression equation for this solution is as follows: 
(2) Log X 12 = -0.09406 log Xj -0.52655 log X 3 +0.30832 log X 5 
+ 2.60325 logX 6 -0.33002 log X 7 +0.3546 log X 8 +0.02428 X 9 + (K) 
The relative importance of the factors is as follows: 
Per cent 
Xi Monthly slaughter 6. 3 
X 3 Moving average of slaughter. 26. 
X 5 Business activity 5. 7 
X 6 Population of United States 37. 6 
X 7 Price of steers — 13. 1 
X 8 European demand 6. 2 
X 9 Time (changing consumption habits) 5. 6 
Total determination by all factors (R 2 ) 74. 4 
After leaving out the upward trend of prices due to the growth 
of population and the increasing per capita demand (the regressions 
are positive for both X 6 and X 9 ), the supply is the dominant factor in 
determining the price. As shown by both the net regression 
coefficients and the coefficients of determination, the trend of supply 
has a much more important effect upon the price than the supply 
during a single month, a given change in the trend of supply having 
«rj,3=+0.53, n, 4= +0.62. 
